TensorFlow with Jupyter Notebooks using Virtualenv

Been trying to learn TensorFlow by working on the Udacity Deep Learning mooc.  All the programming assignments are based on Jupyter Notebooks.  Unfortunately, since I setup my computer with a NVIDIA GPU I've been using Virtualenv to mange my Python distributions as recommended in the Tensorflow installation documents.  However, I've had a really hard time getting IPython and Jupyter configured so I can access all the packages I needed until I read this.

The solution is quite simple.  From your tensorflow environment, first install ipykernel. Then you register the kernel with the tensorflow environment.

$ source ~/tensorflow/bin/activate
$ pip install ipykernel
$ python -m ipykernel install --user --name=tensorflow

Finally, when you open your notebook you will have to change kernels from the default Python ones to the special tensorflow one.  

Machine Learning For Cocktail Parties

I finally completed the MOOC from Columbia on Machine Learning for Data Science and Analytics and received my verified certificate.  I didn't mention this in my past post, but in the end I did decide to register for the verified course.  I was hoping it would improve.  As you can see, I think it was a total waste.  

This was by far THE worst and THE most expensive MOOC I've ever taken.  There was no math or any programming.  I did learn some information on interesting topics like genetics and depression etc but that's about it.  Nothing that was discussed about machine learning went beyond the superficial.  No a single formula on linear regression or logistic regression.  Not even a discussion on the differences between SVN or Neural Networks.  This was really a waste of my time.  I regret paying $99.